Profiling Initial Thrombin Generation in Cardiovascular Disease Using a High Sensitivity Coagulation Assay

Abstract

Background

Initial thrombin (FIIa) generation is a critical trigger for the amplification and propagation phases of coagulation, driven by two distinct pathways: the tissue factor (TF)-driven and the FVIIIa/FIXa-dependent pathways. However, the clinical utility of measuring initial thrombin generation (ITG) as a marker of thrombogenicity or as a tool to monitor the efficacy of oral anti-FXa therapy remains uncertain.

Methods

ITG driven by TF and the FVIIIa/FIXa complex was first measured in plasma samples from healthy adults (n = 40). This was followed by an analysis of ITG profiles in 771 consecutive patients with cardiovascular diseases to evaluate the effects of anticoagulant therapy and clinical characteristics.

Results

Of the 771 patients studied, 169 were receiving direct oral anticoagulants (DOACs). DOAC treatment significantly suppressed thrombin generation via both TF-driven and FVIIIa/FIXa-dependent pathways. Receiver operating characteristic (ROC) analysis confirmed the strong discriminatory power of both pathways for detecting DOAC use (FVIIIa/FIXa: AUC 0.863, 95% CI: 0.826–0.900; TF: AUC 0.887, 95% CI: 0.856–0.917; both p values < 0.0001). Among patients not on anticoagulants, logistic regression revealed that dialysis was associated with reduced thrombin generation through both pathways. Furthermore, chronic kidney disease and active cancer were specifically associated with diminished TF-driven thrombin generation.

Conclusion

The ITG potentials driven by TF- and FVIIIa/FIXa-dependent pathways represent promising biomarkers for evaluating anticoagulant efficacy. Moreover, the distinct influence of pathological conditions on each pathway underscores the need to account for specific disease contexts when assessing coagulation potential.

Keywords

thrombin formation - tissue factor - factor VIII - factor IX - direct oral anticoagulants - DOACs

Introduction

Anticoagulation therapy for thrombotic disorders has undergone a major paradigm shift in recent years.[1] [2] [3] The introduction of direct oral anticoagulants (DOACs) has transformed clinical practice, challenging the long-standing reliance on warfarin as the standard treatment.[4] [5] In ischemic heart disease, growing evidence supports the efficacy of anticoagulants in secondary prevention, comparable to that of antiplatelet therapy, which has traditionally been the first-line choice for secondary prevention.[6] [7] As a result, the optimal management of antithrombotic strategies is being actively reexamined. Central to this effort is the need for precise and individualized assessment of thrombotic risk to evaluate the effectiveness of antithrombotic agents.[8] Early and sensitive detection of thrombotic tendencies is not only crucial for optimizing treatment but also for preventing thrombotic events, including ischemic cardiovascular complications.

In the current model of the coagulation cascade, the initiation phase is triggered by the tissue factor (TF)–FVIIa complex, which activates factor X (FXa), leading to the formation of the prothrombinase complex with the active cofactor FVa and the generation of small amounts of initial thrombin (FIIa) via activation of prothrombin. This initial FIIa acts in a feedback loop to amplify the prothrombinase complex, culminating in a rapid thrombin burst essential for fibrin clot formation.[9] [10] [11] Beyond the TF pathway, recent findings suggest that TF can also initiate thrombin generation through the FVIIIa/FIXa complex coagulation pathway by directly activating FVIII and FIX during the initiation phase.[12]

Detecting initial thrombin generation (ITG) offers a potentially powerful marker for early thrombotic activity, yet measuring such small thrombin quantities has posed a technical challenge. To address this, we recently developed a highly sensitive assay capable of independently quantifying ITG in plasma via the TF-driven and FVIIIa/FIXa-dependent pathways.[12] Using this assay, we demonstrated that TF-driven thrombin generation is elevated in obese patients, underscoring the clinical relevance of these measurements.[13]

However, the applicability of this sensitive thrombin generation assay in patients with cardiovascular diseases and patients already receiving antithrombotic therapy remains unexplored. In this study, we sought to evaluate this highly sensitive thrombin generation assay in a cohort of patients with cardiovascular disease to better understand the usefulness of this measurement for the evaluation and treatment of cardiovascular diseases.

Materials and Methods

Materials

The following reagents were obtained from commercial suppliers: Human TF (Dade Innovin), control plasma N, and FVIII- and FIX-deficient plasma (Sysmex, Kobe, Japan); procoagulant phospholipids (PLs) including 70% di-oleyl phosphatidylcholine/30% di-oleyl phosphatidylserine (mol/mol) (Haematex, Hornsby, NSW, Australia); direct anti-FXa drugs rivaroxaban, edoxaban, and apixaban (ChemScene, Monmouth Junction, NJ, USA); anti-human FVIII monoclonal antibody (MoAb; B2), which inhibits FVIII activity (Life Sciences Research Partners, Leuven, Belgium); human FXIa and corn trypsin inhibitor (CTI) (Prolytix, Essex Junction, VT, USA).

Study Population

Plasma Samples from Healthy Adults

Plasma samples were collected from healthy adult volunteers to evaluate ITG. The study was approved by the Institutional Review Board of Maebashi Kita Hospital (Maebashi, Gunma, Japan). All participants (n = 40; age range: 48–63 years; males: n = 20; females: n = 20) provided written, informed consent prior to enrollment. Whole blood was drawn from the antecubital vein into tubes containing 3.2% trisodium citrate. Platelet-poor plasma (PPP) was prepared by centrifugation at 2,500 × g for 10 minutes at 25°C.

Plasma Samples from Patients with Cardiovascular Diseases

This cross-sectional study included 771 consecutive patients admitted to Kumamoto University Hospital between January 2022 and March 2023 for the diagnosis or treatment of cardiovascular diseases. Patients who received heparin prior to blood collection were excluded (n = 9). Whole blood was collected during catheter examinations into 3.2% trisodium citrate. PPP was obtained by centrifugation at 2,500 × g for 10 minutes at 25°C and stored at −80°C until analysis.

Of the 762 enrolled patients, 169 patients were receiving anticoagulant therapy at the time of blood collection, including DOACs (n = 140) and warfarin (n = 29). The second analysis was performed on 593 patients who were not on any anticoagulant medication ([Fig. 1]).

Fig. 1 Patient recruitment process. A total of 771 consecutive patients who underwent cardiac catheterization between January 2022 and March 2023 were enrolled. Nine patients were excluded because of heparin contamination, and the remaining 762 patients (470 males and 292 females) were divided into two groups (non-anticoagulants, n = 593; anticoagulants, n = 169). Of the 169 receiving anticoagulant therapy, 140 were treated with direct oral anticoagulants (DOACs: rivaroxaban, n = 50; edoxaban, n = 50; apixaban, n = 29) and 29 with warfarin.

All study procedures were conducted in accordance with the Declaration of Helsinki and its amendments. The collection and storage of patient samples were approved by the Human Ethics Committee of Kumamoto University (Approval No. 472). In addition, the use of the samples for this study was approved by the Human Ethics Committee of Kumamoto University (Approval No. 2492). Informed consent was obtained either in writing or via an opt-out procedure, as approved by the ethics board and disclosed to participants.

Assay of Initial Thrombin Generation (ITG) in Plasma

The TF-driven ITG in human plasma was induced by adding 10 µL of a reagent including TF (2.5 pm), anti-FVIII MoAb (0.6 µg/mL) to inhibit FVIIIa activity, and PLs (20 µM) into 45 µL of recalcified plasma containing 16 mM CaCl₂ in microtiter plate wells, as previously described.[13] After 2.5 minutes incubation at 37°C, the ITG reaction was stopped by adding ethylenediaminetetraacetic acid (EDTA). Thrombin (FIIa) activity was measured kinetically by monitoring the fluorescence intensity generated by the cleavage of the fluorogenic substrate, D-cyclohexylalanyl-alanyl-arginine-7-amido-4-methylcoumarin (Pentapharm AG, Aesch BL, Switzerland), at a final concentration of 50 µM. Fluorescence intensity was recorded over 2 minutes at 37°C using a microplate reader, with excitation and emission wavelengths of 355 and 460 nm, respectively. The rate of fluorescence increase was converted into FIIa-equivalent concentrations using a standard calibration curve constructed with a FIIa calibrator (DiaPharma Group Inc., West Chester Township, OH, USA). The levels were also expressed as a relative ratio (%) of that in pooled control plasma N (Sysmex).

To evaluate FVIIIa/FIXa-dependent ITG, we used a modified assay incorporating a reagent mixture containing TF (0.15 pm), FXIa (75 pm), and PLs (20 µM) in recalcified plasma as described above. The dose–response relationship of FVIII and FIX (ranging from 0 to 94%) was determined using plasma samples prepared by diluting normal control plasma with FVIII- or FIX-deficient plasma to achieve concentrations ranging from 0 to 94%. ITG increased in a dose-dependent manner with increasing concentrations of FVIII and FIX ([Supplementary Fig. S1A], available in the online version only). Importantly, addition of CTI (20 µg/mL), which inhibits FXIIa, had no effect on ITG in plasma from healthy volunteers, confirming the absence of contact activation in ITG ([Supplementary Fig. S1B], available in the online version only). The intra-assay and inter-assay coefficients of variation (CVs) for the ITG assay using control plasma N were both below 10%.

To investigate the effects of direct anti-FXa drugs on ITG, plasma samples from patients not receiving anticoagulants were spiked with rivaroxaban, edoxaban, or apixaban. These samples were then analyzed using the ITG assay as described above.

Continuous TG Assay of the Total Amount of FIIa Generated in Plasma

We determined the total amount of FIIa generated in plasma using the continuous TG assay as previously described.[12] [13] The TF-dependent TG was induced by adding TF (2.5 pm) and PLs (20 µM) concomitantly with 18 mM CaCL₂ into plasma (53 µL) containing 73 µM benzyloxycarbonyl-glycyl-glycyl-l-arginine coupled to 7-amido-4-methylcoumarin (Peptide Institute Inc., Osaka, Japan) as a FIIa substrate, followed by kinetically monitoring FIIa activity at 37°C for 40 minutes. Based on the resulting thrombograms, we calculated the TG parameters peak FIIa levels and endogenous thrombin potential (ETP) using GraphPad Prism software (version 9.5.1; GraphPad Software, San Diego, CA, USA).

To validate the ITG assay, we analyzed the correlation between the levels of TF-driven ITG and the parameters obtained from the continuous TF-induced TG assay. For this, we utilized plasma samples including various concentrations of anti-FXa drug (30, 60, 120, 240, and approximately 460–480 ng/mL), which were expected to exhibit altered coagulation potential. We observed a significant correlation of the TF-driven ITG levels with peak FIIa levels (r = 0.926, p < 0.0001, [Supplementary Fig. S2A], available in the online version only) and ETP (r = 0.789, p < 0.001, [Supplementary Fig. S2B], available in the online version only). These results suggest that the ITG is a key determinant of overall coagulation potential.

Statistical Analysis

Normally distributed data are presented as mean ± standard deviation (SD), whereas non-normally distributed data are reported as medians. Categorical variables are presented as frequencies and percentages. Group comparisons were performed using the chi-square (χ²) test for categorical variables and one-way analysis of variance (ANOVA) for continuous variables, as appropriate. Pearson's correlation coefficient was used to evaluate the association between parameters of thrombin generation pathways. Survival analysis was performed using the Kaplan-Meier method, and differences between groups were assessed by the log-rank test. To examine the associations between patient clinical characteristics and thrombin generation, multiple logistic regression analysis was performed using the forced entry method. All statistical analyses were conducted using SPSS version 26 (SPSS Inc., Tokyo, Japan). Standard model diagnostics indicated an adequate fit of the multivariable logistic regression model.

Results

Profiles of FVIIIa/FIXa-dependent and TF-driven Initial Thrombin Generation (ITG) in Healthy Adults

We first characterized the ITG profiles in plasma samples from 40 healthy adults (n = 40; 20 males and 20 females). FVIIIa/FIXa-dependent ITG showed a median of 62.2% relative to pooled control plasma, with no significant sex-based difference (males: median 66.3%, interquartile range [IQR] 53.5–78.9%; females: median 55.7%, IQR 42.5–79.2%) ([Fig. 2A]). TF-driven ITG also showed no significant difference between sex (males: median 74.9%, IQR 67.9–84.4%; females: median 76.4%, IQR 68.0–88.0%] ([Fig. 2B]).

Fig. 2 The percentage of FVIIIa/FIXa-dependent and tissue factor (TF)-driven initial thrombin generation (ITG) in healthy adults. (A) FVIIIa/FIXa-dependent ITG in plasma from 20 males and 20 females. (B) TF-driven initial thrombin generation in 20 males and 20 females. The values are expressed as the percentages relative to pooled control plasma (Control plasma N; Sysmex, Kobe, Japan). Differences between groups (shown as 25th–75th percentile ranges, whiskers indicating minimum-to-maximum values, and a line at the median) were analyzed using Mann-Whitney test.

FVIIIa/FIXa-dependent and TF-driven ITG in Cardiovascular Disease Patients

To assess the assay's ability to detect the effect of DOAC therapy on thrombin generation, plasma from patients (n = 20) with cardiovascular diseases was spiked with the direct FXa inhibitors, rivaroxaban, edoxaban, or apixaban. All three DOACs suppressed FVIIIa/FIXa-dependent and TF-driven ITG in a dose-dependent manner, with TF-driven ITG showing greater sensitivity to inhibition ([Fig. 3A, B]). These findings indicate that the assay can differentiate between thrombin generation pathways and detect DOAC effects with high sensitivity.

Fig. 3 Effect of direct anti-FXa drugs on initial thrombin generation (ITG). We tested the effect of direct anti-FXa drugs (rivaroxaban, edoxaban, and apixaban) on ITG in plasma from patients (n = 20). (A) Inhibition of tissue factor (TF)-driven ITG by the drugs at indicated concentrations. The ITG was induced by incubating plasma with TF (2.5 pm)/anti-FVIII monoclonal antibody (MoAb)/procoagulant phospholipids (PLs). (B) Inhibition of FVIIIa/FIXa-dependent ITG by the drugs at the indicated concentrations. The ITG was induced by incubating plasma with TF (0.15 pm)/FXIa (75 pm)/PLs. Difference between groups (shown as 25th–75th percentile bars, min-to-max whiskers and median line) were analyzed using one-way ANOVA tests. ****p < 0.0001.

Next, to evaluate the clinical applicability of this assay, we analyzed plasma samples from 762 patients with cardiovascular diseases who were receiving anticoagulation therapy. The mean age was 72.1 ± 12.0 years, with 470 individuals (61.7%) being male. Among them, 169 patients were receiving anticoagulant therapy, including 140 on DOACs (rivaroxaban = 50, apixaban = 29, edoxaban = 56) and 29 on warfarin ([Fig. 1]).

Compared with patients not receiving anticoagulant therapy (n = 593), those on anticoagulants were older (74.7 ± 11.1 vs. 71.3 ± 12.1 years, p < 0.001) and had higher prevalence of atrial fibrillation (66.9% vs. 4.6%, p < 0.001). In contrast, they had lower rates of dyslipidemia (47.3% vs. 60.9%, p = 0.001), diabetes mellitus (28.4% vs. 39.6%, p = 0.009), ischemic heart disease (6.5% vs. 22.8%, p < 0.001), currently smoking (4.7% vs. 17.7%, p < 0.001), and active cancer (3.6% vs. 9.1%, p = 0.022) ([Table 1]). D-dimer levels were also significantly lower in the group receiving anticoagulant therapy (median 0.40 µg/mL [IQR: 0.30–0.78] vs. 0.85 [0.50–2.43], p < 0.001] ([Table 1]).

Table 1 Comparison of baseline demographic and clinical parameters between patients with and without anticoagulants

Baseline character	With anticoagulants (n = 169)	Without anticoagulants (n = 593)	P value
Mean age (years)	74.7 ± 11.1	71.3 ± 12.1	0.001
Male (%)	103 (60.9)	367 (61.9)	0.858
Mean body mass index (kg/m2)	23.3 ± 4.6	23.2 ± 4.1	0.654
Hypertension (%)	124 (73.4)	464 (78.2)	0.212
Dyslipidemia (%)	80 (47.3)	361 (60.9)	0.001
Diabetes (%)	48 (28.4)	235 (39.6)	0.009
Stroke (%)	17 (10.1)	53 (8.9)	0.656
Atrial fibrillation (%)	113 (66.9)	27 (4.6)	<0.001
OMI (%)	11 (6.5)	135 (22.8)	<0.001
Current smoking (%)	8 (4.7)	105 (17.7)	<0.001
Cancer (%)	6 (3.6)	54 (9.1)	0.022
History of cancer (%)	23 (13.6)	111 (18.7)	0.137
eGFR	53.1 ± 22.0	56.2 ± 34.5	0.268
T-Cho (mg/dL)	172.8 ± 40.4	178.7 ± 43.6	0.137
TG (mg/dL)	107.8 ± 50.4	133.5 ± 86.0	<0.001
HDL-Cho (mg/dL)	58.1 ± 15.5	56.8 ± 16.9	0.410
LDL-Cho (mg/dL)	96.0 ± 32.3	96.9 ± 35.5	0.771
WBC (103/μL)	6.1 ± 2.2	6.7 ± 3.0	0.020
Hgb (g/dL)	12.7 ± 2.5	12.9 ± 4.4	0.643
Hct (%)	40.0 ± 7.8	39.1 ± 6.2	0.182
Platelet count (103/μL)	203.9 ± 73.6	210.8 ± 69.4	0.260
D-dimer (μg/mL)	0.40 [0.30, 0.78]	0.85 [0.50, 2.43]	<0.001
PT-INR	1.57 [1.24, 2.00]	0.98 [0.93, 1.04]	<0.001
FVIIIa/FIXa-dependent initial thrombin generation value (%)	2.54 [0.87, 11.4]	54.3 [29.8, 74.2]	<0.001
TF-driven initial thrombin generation value (%)	1.72 [0.66, 8.0]	41.4 [22.3, 62.7]	<0.001

Abbreviations: eGFR, estimated glomerular filtration rate; Hgb, hemoglobin; Hct, hematocrit; HDL-Cho, high-density lipoprotein cholesterol; LDL-Cho, low-density lipoprotein cholesterol; OMI, old myocardial infarction; T-Cho, total cholesterol; TG, triglyceride; WBC, white blood cell count.

Note: Data are mean ± SD or n (%).

Across the total patient cohort, FVIIIa/FIXa-dependent ITG had a median of 44.8% (IQR: 10.6–70.5%), and TF-driven ITG had a median of 33.4% (IQR: 8.4–58.6%), both notably reduced compared with pooled control plasma ([Fig. 4A, B]). A moderate correlation was observed between the two assays [r = 0.570, p < 0.0001] ([Fig. 4C]).

Fig. 4 FVIIIa/FIXa-dependent and tissue factor (TF)-driven initial thrombin generation (ITG) in patients with cardiovascular diseases. (A) FVIIIa/FIXa-dependent ITG in plasma from patients with cardiovascular diseases (n = 762). (B) TF-driven ITG in plasma from patients with cardiovascular diseases (n = 762). The values are expressed as percentages relative to pooled control plasma (Control plasma N; Sysmex, Kobe, Japan). Each box-and-whisker plot represents the median value with 25th–75th percentile ranges and whiskers indicating minimum-to-maximum values. (C) Scatter plot showing the correlation between FVIIIa/FIXa-dependent and TF-driven ITG in plasma (r = 0.570, p < 0.0001; Pearson's correlation coefficient).

We next compared PT-INR and thrombin generation profiles between patients receiving anticoagulant therapy and those not on anticoagulants. PT-INR values were significantly higher in patients receiving anticoagulant therapy (1.57 [1.24–2.00]) compared with the non-anticoagulated group (0.98 [0.93–1.04]; p < 0.001) ([Table 1]). Both FVIIIa/FIXa-dependent and TF-driven ITG were substantially suppressed in patients on anticoagulants. Specifically, FVIIIa/FIXa-dependent ITG was reduced to 2.54 (0.87–11.4) compared with 54.3 (29.8–74.2) in the non-anticoagulated group (p < 0.001), and TF-driven ITG was 1.72 (0.66–8.0) versus 41.4 (22.3–62.7), respectively (p < 0.001).

FVIIIa/FIXa-dependent and TF-driven ITG Accurately Detect the Effects of DOACs

Receiver operating characteristic (ROC) curve analyses were performed to assess the diagnostic performance of thrombin generation assays in detecting DOAC use. The area under the curve (AUC) for FVIIIa/FIXa-dependent ITG was 0.863 (95% confidence interval (CI) 0.826–0.900, p < 0.0001), and that for TF-driven ITG was 0.887 (95% CI 0.856–0.917, p < 0.0001) ([Fig. 5A, B]). Subgroup analyses for individual DOACs showed AUCs > 0.800 for FVIIIa/FIXa-dependent ITG ([Fig. 5C, E, G]) and >0.850 for TF-driven ITG ([Fig. 5D, F, H]), confirming consistent performance across different anticoagulant agents.

Fig. 5 Receiver operating characteristic (ROC) curves of FVIIIa/FIXa-dependent and tissue factor (TF)-driven initial thrombin generation (ITG) for detecting direct oral anticoagulant (DOAC) treatment. (A, B) Overall ROC curve analysis of FVIIIa/FIXa-dependent and TF-driven ITG for detecting any DOAC treatment (DOAC users, n = 140; non-anticoagulants, n = 593). (C–H) Subgroup ROC analyses for each DOAC: rivaroxaban (C and D, n = 50), edoxaban (E and F, n = 56), and apixaban (G and H, n = 29), each compared with non-anticoagulants (n = 593).

Furthermore, ROC curve analyses were also performed in patients receiving warfarin (n = 29), which showed high AUCs (>0.900 for FVIIIa/FIXa-dependent ITG and >0.850 for TF-driven ITG) ([Supplementary Fig. S3A, B], available in the online version only), comparable to those observed with DOACs.

These findings demonstrate the strong potential of both FVIIIa/FIXa-dependent and TF-driven ITG as sensitive biomarkers for assessing anticoagulant effects in clinical settings.

Impact of Pathological Conditions on FVIIIa/FIXa-dependent and TF-driven ITG

Based on the findings thus far, FVIIIa/FIXa-dependent and TF-driven ITG values can serve as reliable indicators for assessing the pharmacological effects of anticoagulant therapy. To investigate the influence of comorbid conditions, we analyzed the 593 patients not on anticoagulant therapy. In this subgroup, FVIIIa/FIXa-dependent ITG remained reduced (median: 54.3%; IQR: 29.8–74.2%) compared with pooled control plasma, as did TF-driven ITG (median: 41.4%; IQR: 22.3–62.7%) ([Fig. 6A, B]). In this subgroup, the correlation between assays was weaker (r = 0.390, p < 0.0001) compared with the full cohort that included patients receiving anticoagulant therapy ([Fig. 6C] and [Fig. 4C]).

Fig. 6 FVIIIa/FIXa-dependent and tissue factor (TF)-driven initial thrombin generation (ITG) in patients not receiving anticoagulant therapy. (A) FVIIIa/FIXa-dependent ITG in plasma from patients not receiving anticoagulant therapy (n = 593). (B) TF-driven ITG in patients not receiving anticoagulant therapy (n = 593). The values are expressed as the percentages relative to pooled control plasma (Control plasma N; Sysmex). Each box-and-whisker plot represents the median value with 25th–75th percentile ranges and whiskers indicating minimum-to-maximum values. (C) Scatter plot showing the correlation between FVIIIa/FIXa-dependent and TF-driven ITG in plasma (r = 0.390, p < 0.0001; Pearson's correlation coefficient).

These results suggest that patients with cardiovascular diseases tend to exhibit reduced FVIIIa/FIXa-dependent and TF-driven ITG compared with pooled control plasma. To further investigate factors associated with lower thrombin generation, each parameter was dichotomized at its median value, and univariate and multivariate logistic regression analyses were performed. Univariate logistic regression analysis revealed that diabetes mellitus (odds ratio [OR] 1.453, 95% CI 1.046–2.018, p = 0.026), dyslipidemia (OR 1.471, 95% CI 1.056–2.050, p = 0.022), and dialysis treatment (OR 4.165, 95% CI 2.034–8.528, p < 0.001) were significant factors associated with lower FVIIIa/FIXa-dependent ITG. In addition, statin use was associated with lower FVIIIa/FIXa-dependent ITG (OR 1.340, 95% CI 0.968–7.855, p = 0.078), indicating a trend toward significance.

In multivariate logistic regression analysis, older age (OR 0.753, 95% CI 0.532–1.065, p = 0.190), male gender (OR 0.970, 95% CI 0.687–1.370, p = 0.862), diabetes mellitus (OR 1.173, 95% CI 0.827–1.663, p = 0.372), dyslipidemia (OR 1.342, 95% CI 0.900–2.002, p = 0.149), and statin use (OR 1.102, 95% CI 0.743–1.636, p = 0.631) were not significantly associated with FVIIIa/FIXa-dependent ITG. In contrast, dialysis treatment (OR 3.708, 95% CI 1.789–7.727, p < 0.001) remained an independent predictor of reduced FVIIIa/FIXa-dependent ITG ([Table 2]).

Table 2 Univariate and multivariate logistic regression analyses of factors predicting low FVIIIa/FIXa-dependent ITG (<54.3)

	Univariate			Multivariate
	OR	95% CI	P value	OR	95% CI	P value
Age (>75 years)	0.699	0.501–0.974	0.034	0.753	0.532–1.065	0.109
Male (yes)	1.018	0.735–1.410	0.916	0.970	0.687–1.370	0.862
Body mass index (>25 kg/m2)	0.858	0.600–1.227	0.402
Hypertension (yes)	1.204	0.819–1.771	0.344
Diabetes (yes)	1.453	1.046–2.018	0.026	1.173	0.827–1.663	0.372
Dyslipidemia (yes)	1.471	1.056–2.050	0.022	1.342	0.900–2.002	0.149
IHD (yes)	0.930	0.707–1.224	0.606
Stroke (yes)	0.949	0.540–1.668	0.855
PAD (yes)	1.026	0.604–1.742	0.926
CKD (yes)	1.019	0.737–1.408	0.912
Hemodialysis (yes)	4.165	2.034–8.528	<0.001	3.708	1.780–7.727	<0.001
Current smoker (yes)	1.414	0.920–2.172	0.114
Cancer (yes)	1.107	0.646–1.897	0.711	1.130	0.649–7.968	0.665
History of cancer (yes)	1.196	0.798–1.795	0.386
Statin	1.340	0.968–7.855	0.078	1.102	0.743–1.636	0.631
Clopidogrel	1.231	0.753–2.013	0.408
Prasugrel	1.290	0.629–2.645	0.486
Aspirin	1.128	0.814–1.564	0.468
Cilostazol	1.429	0.536–3.805	0.475

Abbreviations: CKD, chronic kidney disease; IHD, ischemic heart disease; ITG, initial thrombin generation; OR, odds ratio; PAD, peripheral artery disease; 95% CI, 95% confidence interval.

Note: Data of this parameter were measured at admission.

Logistic regression analysis was also performed for TF-driven ITG. In the univariate analysis, age (OR 0.923, 95% CI 0.664–1.285, p = 0.636), male gender (OR 1.022, 95% CI 0.738–1.416, p = 0.896), and dyslipidemia (OR 1.359, 95% CI 0.976–1.892, p = 0.069) were not significantly associated with TF-driven ITG. In contrast, dialysis treatment (OR 3.288, 95% CI 1.675–6.456, p = 0.001), active cancer (OR 1.786, 95% CI 1.026–3.108, p = 0.040), chronic kidney disease (CKD) (OR 1.721, 95% CI 1.241–2.386, p = 0.001), and aspirin use (OR 1.614, 95% CI 1.162–2.224, p = 0.004) were significantly associated with reduced TF-driven ITG.

In the multivariate model, CKD (OR 1.490, 95% CI 1.031–2.096, p = 0.033), dialysis treatment (OR 2.634, 95% CI 1.295–5.358, p = 0.008), and active cancer (OR 1.811, 95% CI 1.027–3.195, p = 0.040) remained significant independent predictors of reduced TF-driven ITG, whereas aspirin use was no longer significant ([Table 3]).

Table 3 Univariate and multivariate logistic regression analyses of factors predicting low TF-driven ITG value (<41.4%)

	Univariate			Multivariate
	OR	95% CI	P value	OR	95% CI	P value
Age (>75 years)	0.923	0.664–1.285	0.636	0.534	0.624–1.277	0.635
Male (yes)	1.022	0.738–1.416	0.896	0.958	0.677–1.355	0.808
Body mass index (>25 kg/m2)	1.044	0.731–1.492	0.813
Hypertension (yes)	1.231	0.837–1.811	0.290
Diabetes (yes)	1.034	0.746–1.433	0.842
Dyslipidemia (yes)	1.359	0.976–1.892	0.069	1.208	0.852–1.713	0.289
IHD (yes)	1.030	0.789–1.344	0.827
Stroke (yes)	0.750	0.425–1.324	0.321
PAD (yes)	1.398	0.818–2.387	0.220
CKD (yes)	1.721	1.241–2.386	0.001	1.470	1.031–2.096	0.033
Hemodialysis (yes)	3.288	1.675–6.456	0.001	2.634	1.295–5.358	0.008
Current smoker (yes)	1.148	0.749–1.760	0.526
Cancer (yes)	1.786	1.026–3.108	0.040	1.811	1.027–3.195	0.040
History of cancer (yes)	1.117	0.745–1.675	0.592
Statin	1.140	0.824–1.577	0.429
Clopidogrel	0.973	0.596–1.588	0.913
Prasugrel	1.309	0.639–2.683	0.462
Aspirin	1.614	1.162–2.241	0.004	1.415	0.996–2.012	0.053
Cilostazol	1.449	0.544–3.858	0.458

Abbreviations: CKD, chronic kidney disease; IHD, ischemic heart disease; ITG, initial thrombin generation; OR, odds ratio; PAD, peripheral artery disease; TF, tissue factor; 95% CI, 95% confidence interval.

Note: Data of this parameter were measured at admission.

Association Between the ITG and Prognosis in Patients not Receiving Anticoagulants

We assessed mortality rate as the endpoint to compare the prognosis implications for FVIIIa/FIXa-dependent and TF-driven ITG values, respectively. The mean observation period was 645.8 days. Kaplan–Meier analysis was performed using the date of hospital admission as the starting point. Among patients who were not treated with warfarin or DOACs, there was no significant difference in survival between the high and low FVIIIa/FIXa-dependent ITG groups (Log-rank test: χ² = 0.9206, p = 0.3373) ([Fig. 7A]). In contrast, analysis based on TF-driven ITG showed that, although the difference did not reach statistical significance, patients in the low TF group tended to have lower survival (log-rank test: χ² = 3.224, p = 0.0726) ([Fig. 7B]). These results indicate that FVIIIa/FIXa-dependent ITG was not associated with survival outcome, whereas TF-dependent ITG, although not statistically significant, suggested a potential association with mortality.

Fig. 7 Kaplan-Meier survival curves stratified by thrombin generation in patients not receiving anticoagulant therapy. (A) Survival curves according to FVIIIa/FIXa-dependent initial thrombin generation (ITG). Patients were divided into high and low FVIIIa/FIXa-dependent ITG groups. A total of 19 and 21 deaths were observed in the high and low FVIIIa/FIXa groups, respectively (log-rank test, p = 0.3373). (B) Survival curves according to tissue factor (TF)-driven ITG. Patients were divided into high and low TF-driven ITG groups. A total of 14 and 26 deaths were observed in the high and in low TF-driven ITG group, respectively (log-rank test, p = 0.0726).

Discussion

In this study, we evaluated the clinical utility of measuring ITG via FVIIIa/FIXa-dependent and TF-driven pathways. Our initial analysis confirmed that in vitro administration of DOACs resulted in a dose-dependent suppression of thrombin generation. In subsequent analysis of patients undergoing anticoagulant therapy, both assays reliably detected the presence of DOAC therapy with high sensitivity. Beyond anticoagulant monitoring, we also investigated how underlying clinical conditions affect thrombin generation. Dialysis treatment was independently associated with reduced FVIIIa/FIXa-dependent ITG, while CKD, dialysis, and active cancer were significantly associated with reduced TF-driven ITG. To our knowledge, this is the first study to demonstrate that measuring ITG in clinical samples not only reflects anticoagulant efficacy but also provides meaningful insights into underlying pathophysiological states.

A key advantage of DOACs is that they can be administered without the intensive monitoring required for warfarin therapy.[14] However, in elderly patients, the anticoagulant therapeutic window is narrow, and in those with multiple comorbidities or polypharmacy, potential drug interactions must be carefully considered.[15] [16] This underscores the importance of accurately assessing DOAC effectiveness to support personalized anticoagulation strategies. In this study, both the FVIIIa/FIXa-dependent and TF-driven ITG assays demonstrated a marked decrease in thrombin generation following DOAC administration, with levels approaching complete suppression relative to control plasma. These results confirm that current dosing of DOACs robustly inhibits the initial phase of thrombin generation. Not only do these results reinforce previous evidence of DOAC efficacy,[17] but they also highlight ITG as an important indicator for future dose optimization.

Among patients not receiving anticoagulant therapy, our analysis revealed both shared and distinct factors influencing FVIIIa/FIXa-dependent and TF-driven ITG. A key common finding was that dialysis was consistently associated with reduced ITG in both assays. This is notable given that dialysis patients are known to have a paradoxical risk of both bleeding and thromboembolic complications.[18] Prior studies using automated thrombin generation assays (thrombogram) have shown conflicting results: patients undergoing dialysis with shunt occlusion exhibited elevated thrombin generation, whereas stable dialysis patients had reduced endogenous thrombin potential compared with healthy individuals.[19] These observations underscore the complex, dual nature of coagulation disturbances in this population. Interestingly, CKD was associated with lower TF-driven ITG. Although previous studies of CKD have reported increased thrombotic potential using thromboelastography and continuous TG assay, they have also demonstrated elevated levels of tissue factor pathway inhibitor (TFPI), a key inhibitor of TF-mediated coagulation.[20]

Interestingly, the presence of active cancer was also associated with reduced TF-driven ITG. Although it has long been recognized that cancer is associated with increased thrombotic risk[21] [22] and some tumor cells are known to express TF, thereby promoting TF-driven thrombin generation,[23] our findings present a paradox. Contrary to expectations, a significant reduction was observed specifically in the TF-driven pathway. This unexpected finding suggests a more complex regulatory mechanism in vivo. TFPI has been reported to be upregulated alongside TF in certain tumors, and elevated TFPI levels have also been associated with poor prognosis in cancer patients.[24] [25] [26] Moreover, increased circulating TFPI has been observed in patients with solid malignancies, further supporting the notion that TFPI overexpression may contribute to suppression of TF-dependent coagulation in cancer.[27] In our study, the reduction in thrombin generation was confined to the TF-driven pathway and was not observed in the FVIIIa/FIXa-dependent pathway, indicating a selective disruption of the TF-dependent coagulation cascade in the context of cancer.

In addition, aspirin use was also significantly associated with reduced TF-driven ITG in the univariate analysis (OR 1.614, 95% CI 0.544–3.858, p = 0.004), which is consistent with previous reports demonstrating that aspirin attenuates thrombin generation in a dose-dependent manner, particularly at higher doses.[28]

Furthermore, in this study, diabetes mellitus and dyslipidemia were associated with lower FVIIIa/FIXa-dependent thrombin generation in the univariate analysis, whereas TF-driven thrombin generation did not differ significantly between groups. These findings suggest that metabolic risk factors may differentially influence the intrinsic and extrinsic pathways of coagulation. Such changes may reflect a complex imbalance between procoagulant and anticoagulant mechanisms rather than a simple shift toward hypercoagulability, thereby highlighting the multifaceted nature of thrombogenesis under atherosclerotic conditions.

This study has several limitations. First, the study was conducted at a single center with a relatively small sample size, limiting our ability to fully evaluate the specific effects of DOAC or changes related to disease state. Larger, multi-center studies are needed to validate these findings. Additionally, all participants had underlying cardiovascular disease, and their comorbidities and medication profiles may have influenced thrombin generation results. Furthermore, TF-driven ITG showed a trend toward an association with mortality ([Fig. 7B]); therefore, large-scale and prospective studies are needed to determine whether TF-driven ITG has a prognostic role. To more accurately evaluate intrinsic ITG, future studies should include a well-defined adult population and adopt a prospective design.

In conclusion, this study demonstrates that FVIIIa/FIXa-dependent and TF-driven ITG assays are promising and reliable markers for assessing coagulability in patients receiving various anticoagulation therapies. Beyond anticoagulation monitoring, each assay also shows potential as an early indicator of underlying pathological conditions. Further validation in diverse patient populations with different cardiovascular and systemic diseases is essential to establish their broader clinical utility. Given its ability to sensitively detect early thrombin formation, this high-sensitivity ITG assay may also serve as a potential tool for clinical monitoring and individualized antithrombotic management.

Conflict of Interest

This research is a collaborative study with Thrombo Translational Research Labo Inc. (Kumamoto, Japan) and was funded by TAUNS Laboratories, Inc. (Shizuoka, Japan). The sponsor had no role in the study design and conduct, data collection and analysis, decision to publish, or manuscript preparation.

Acknowledgment

We thank all the paramedical staff and clinical secretaries for their kind support during this work.

• References

• 1 Ajjan RA, Kietsiriroje N, Badimon L. et al. Antithrombotic therapy in diabetes: which, when, and for how long?. Eur Heart J 2021; 42 (23) 2235-2259
• 2 Martens KL, Dekker SE, Crowe M, DeLoughery TG, Shatzel JJ. Challenging clinical scenarios for therapeutic anticoagulation: a practical approach. Thromb Res 2022; 218: 72-82
• 3 Masunaga N, Ishii M, Oka K. et al; Fushimi AF Registry Investigators. 10-year trends of antithrombotic therapy status and clinical outcomes in patients with atrial fibrillation and renal dysfunction—the Fushimi AF Registry. Circ J 2025; 89 (02) 174-183
• 4 Ruf W. Roles of factor Xa beyond coagulation. J Thromb Thrombolysis 2021; 52 (02) 391-396
• 5 Kitahara H, Yamazaki T, Hiraga T. et al. Impact of underdosing of direct oral anticoagulants on clinical outcomes in patients with atrial fibrillation undergoing percutaneous coronary intervention. Circ J 2025; 89 (02) 195-203
• 6 Yasuda S, Kaikita K, Akao M. et al; AFIRE Investigators. Antithrombotic therapy for atrial fibrillation with stable coronary disease. N Engl J Med 2019; 381 (12) 1103-1113
• 7 Rodriguez F, Harrington RA. Management of antithrombotic therapy after acute coronary syndromes. N Engl J Med 2021; 384 (05) 452-460
• 8 Shaw JR, Castellucci LA, Siegal D, Carrier M. DOAC-associated bleeding, hemostatic strategies, and thrombin generation assays—a review of the literature. J Thromb Haemost 2023; 21 (03) 433-452
• 9 Negrier C, Shima M, Hoffman M. The central role of thrombin in bleeding disorders. Blood Rev 2019; 38: 100582
• 10 Sidonio Jr RF, Hoffman M, Kenet G, Dargaud Y. Thrombin generation and implications for hemophilia therapies: a narrative review. Res Pract Thromb Haemost 2022; 7 (01) 100018
• 11 Hoffman M, Monroe DM. Impact of non-vitamin K antagonist oral anticoagulants from a basic science perspective. Arterioscler Thromb Vasc Biol 2017; 37 (10) 1812-1818
• 12 Kamikubo Y, Mendolicchio GL, Zampolli A. et al. Selective factor VIII activation by the tissue factor-factor VIIa-factor Xa complex. Blood 2017; 130 (14) 1661-1670
• 13 Kamikubo Y, Nagaya S, Inoue R. et al. Tissue factor pathway-driven initial thrombin generation is associated with hypercoagulability in obesity. Thromb Haemost 2025; . Epub ahead of print
• 14 Bhandari M, Pradhan A, Vishwakarma P. et al. Direct oral anticoagulant use in older adults with atrial fibrillation: challenges and solutions. Eur Cardiol 2025; 20: e03
• 15 Hayes KN, Zhang T, Kim DH. et al. Benefits and harms of standard versus reduced-dose direct oral anticoagulant therapy for older adults with multiple morbidities and atrial fibrillation. J Am Heart Assoc 2023; 12 (21) e029865
• 16 Edwina AE, Dia N, Dreesen E. et al. Insights into the pharmacokinetics and pharmacodynamics of direct oral anticoagulants in older adults with atrial fibrillation: a structured narrative review. Clin Pharmacokinet 2023; 62 (03) 351-373
• 17 Cabral KP, Ansell J. Oral direct factor Xa inhibitors for stroke prevention in atrial fibrillation. Nat Rev Cardiol 2012; 9 (07) 385-391
• 18 Ocak G, Noordzij M, Rookmaaker MB. et al. Mortality due to bleeding, myocardial infarction and stroke in dialysis patients. J Thromb Haemost 2018; 16 (10) 1953-1963
• 19 Jeong JC, Kim J-E, Ryu JW, Joo KW, Kim HK. Plasma haemostatic potential of haemodialysis patients assessed by thrombin generation assay: hypercoagulability in patients with vascular access thrombosis. Thromb Res 2013; 132 (05) 604-609
• 20 Lim HY, Lui B, Tacey M. et al. Global coagulation assays in patients with chronic kidney disease and their role in predicting thrombotic risk. Thromb Res 2023; 226: 127-135
• 21 Wan T, Song J, Zhu D. Cancer-associated venous thromboembolism: a comprehensive review. Thromb J 2025; 23 (01) 35
• 22 Nakano Y, Adachi S, Imai R. et al. Mortality, recurrent thromboembolism and major bleeding in cancer-associated and non-cancer pulmonary embolism patients treated with direct oral anticoagulants. Circ J 2024; 88 (02) 243-250
• 23 Koizume S, Miyagi Y. Tissue factor in cancer-associated thromboembolism: possible mechanisms and clinical applications. Br J Cancer 2022; 127 (12) 2099-2107
• 24 Sletten M, Skogstrøm KB, Lind SM. et al. Elevated TFPI is a prognostic factor in hepatocellular carcinoma: putative role of miR-7-5p and miR-1236-3p. Thromb Res 2024; 241: 109073
• 25 Abu Saadeh F, Norris L, O'Toole S. et al. Tumour expresion of tissue factor and tissue factor pathway inhibitor in ovarian cancer—relationship with venous thrombosis risk. Thromb Res 2013; 132 (05) 627-634
• 26 Englisch C, Moik F, Thaler J. et al. Tissue factor pathway inhibitor is associated with risk of venous thromboembolism and all-cause mortality in patients with cancer. Hamostaseologie 2023; 43: S23-S23
• 27 Iversen N, Lindahl AK, Abildgaard U. Elevated TFPI in malignant disease: relation to cancer type and hypercoagulation. Br J Haematol 1998; 102 (04) 889-895
• 28 Wallén NH, Ladjevardi M. Influence of low- and high-dose aspirin treatment on thrombin generation in whole blood. Thromb Res 1998; 92 (04) 189-194

Publication History

Received: 09 July 2025

Accepted after revision: 19 November 2025

Accepted Manuscript online:
21 November 2025

Article published online:
08 December 2025